Studies On The Analysis Of Consumers' Behavior Characteristics And Identification Of Electricity-Stealing Based On Electricity Marketing Data

Based on the current situation of domestic power system reform and the development of electricity-stealing technology,the importance and urgency of anti-electricity-stealing work is increasing day by day.However,how to accurately and efficiently identify users who steal electricity is the first task of anti-theft management.Based on the data of the electricity consumption information collection system that has been widely used,this paper analyzes the user's current,voltage,power factor,imbalance rate of electricity quantity and other curve conditions with the aid of various algorithms.And then realizes the preliminary identification of power stealing users.Finally the suspected power theft objects are accurately monitored.This paper provided reference and help for anti-stealing electricity work.The following is the main content of this paper.First of all,we consulted literature relevant to this subject.We found that the current research on judging users' electricity stealing behavior was often limited to the horizontal comparison between power users,the accuracy of identifying users who stealing electricity was not very high.Therefore,based on a comparative analysis of multiple basic algorithms,this paper determines the use of analytic hierarchy process and LOF algorithm to establish a electricity user behavior characteristic analysis algorithm based on data collected from electricity consumption information,and combines electricity stealing behavior characteristics to develop a electricity stealing user identification scheme.At the same time,the related algorithms are introduced in detail.Secondly,according to the principle of data outlier detection,a weight stealing identification method based on the weighted LOF algorithm was proposed to implement a comprehensive electricity stealing behavior suspicious assessment for users.Analyze the power indicators for the identified user behavior characteristics,we combine the analytic hierarchy process to analyze and quantify the weight of different electrical indicators,and establish a weighted LOF algorithm,which can avoid detection bias due to uneven distribution of power data density.Effectively reduce the impact of other factors that caused by abnormal electrical parameters.Thus,we could lock the suspected user who steals electricity preliminary and accurately.Lastly,for the tracking analysis of locked suspect users,this paper establishes a tracking and detection model of abnormal users based on BP neural network.Using the mind evolution algorithm to optimize the BP neural network,and based on this,a corresponding identification model of electricity theft is established relied on electricity consumption parameters,At last,the actual application analysis of the model is completed through various specific date.By this,we verified the feasibility of this model.